‘It’s not going to solve all of the problems’: Dan Houser says he’s ‘dabbling’ with AI but it’s ‘not as useful’ as some say

Former Rockstar writer Dan Houser says he’s “dabbling” with AI as he works on his next game, but says it’s not yet quite as revolutionary as some people suggest.

Houser has been promoting his new book A Better Paradise Volume One: An Aftermath, which tells the story of a video game project that goes wrong when the AI they created becomes too powerful.

In an interview with Sunday Brunch, Houser was asked about his next game, which is in development at his company Absurd Ventures. He replied that work started on it 18 months ago, and it’s still “another few years in development”.

Houser was then asked if AI was involved in the game, to which he replied that the story would feature AI characters, and that he was experimenting with using AI in development but doesn’t believe it’s a catch-all solution yet.

“In the story, yeah, there’s lots of AI characters,” he replied. “We are dabbling in using AI, but the truth is a lot of it’s not as useful as some of the companies would have you believe yet. It’s not going to solve all of the problems, you know?”

Houser was asked whether, as someone involved in an industry where AI is being used, he could feel that the pace of AI usage was moving more rapidly.

“In specific areas? Yes,” he said. “In other areas… we have a whole field of areas we need technology for, and AI is great at some of the tasks, and can’t do the other tasks yet. So [some] people, they will claim it can solve every single problem, and it really can’t yet.

“As far as I understand it, it’s a sort of hold-all term for all future computing and it’s really not doing a lot of the stuff yet, but if we all give it all of our money it might do it in the future.”

He then warned that while some aspects of AI are impressive, a lot of it is doing processes “computers are already doing”, adding that “some of it’s just to sell AI stock, or to convince everyone this is transformative, and other stuff it does is amazing.”

“My experience has always been, with technological innovation, that the first 80% is quite easy, and the last 20% – to make it a perfect simulation of something in the real world – is very, very hard. So it’ll be interesting to see that last bit, how hard it is for AI, which learns in a different way to normal computing development, how quick that is or isn’t.”